Discovery of 7-Oxo-2,4,5,7-tetrahydro-6 H -pyrazolo[3,4- c] pyridine Derivatives as Potent, Orally Available, and Brain-Penetrating Receptor Interacting Protein 1 (RIP1) Kinase Inhibitors: Analysis of Structure-Kinetic Relationships

Article abstract of DOI:10.1021/acs.jmedchem.7b01647

We report the discovery of 7-oxo-2,4,5,7-tetrahydro-6H-pyrazolo[3,4-c]pyridine derivatives as a novel class of receptor interacting protein 1 (RIP1) kinase inhibitors. On the basis of the overlay study between HTS hit 10 and GSK2982772 (6) in RIP1 kinase, we designed and synthesized a novel class of RIP1 kinase inhibitor 11 possessing moderate RIP1 kinase inhibitory activity and P-gp mediated efflux. The optimization of the core structure and the exploration of appropriate substituents utilizing SBDD approach led to the discovery of 22, a highly potent, orally available, and brain-penetrating RIP1 kinase inhibitor with excellent PK profiles. Compound 22 significantly suppressed necroptotic cell death both in mouse and human cells. Oral administration of 22 (10 mg/kg, bid) attenuated disease progression in the mouse experimental autoimmune encephalomyelitis (EAE) model of multiple sclerosis (MS). Moreover, analysis of structure-kinetic relationship (SKR) for our novel chemical series was also discussed.

Full text of DOI:10.1021/acs.jmedchem.7b01647

Subscriber access provided by MT ROYAL COLLEGE
Article
Discovery of 7-Oxo-2,4,5,7-tetrahydro-6H-pyrazolo[3,4-c]pyridine Derivatives
as Potent, Orally Available, and Brain-penetrating Receptor Interacting
Protein 1 (RIP1) Kinase Inhibitors; Analysis of Structure–Kinetic Relationships
Masato Yoshikawa, Morihisa Saitoh, Taisuke Katoh, Tomohiro Seki, Simone V. Bigi, Yuji Shimizu,
Tsuyoshi Ishii, Takuro Okai, Masako Kuno, Harumi Hattori, Etsuro Watanabe, Kumar Singh Saikatendu,
Hua Zou, Masanori Nakakariya, Takayuki Tatamiya, Yoshihisa Nakada, and Takatoshi Yogo
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b01647 • Publication Date (Web): 27 Feb 2018
Downloaded from http://pubs.acs.org on February 27, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 110
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery
of
7ꢀOxoꢀ2,4,5,7ꢀtetrahydroꢀ6Hꢀpyrazolo[3,4ꢀc]pyridine
Derivatives as Potent, Orally Available, and
Brainꢀpenetrating Receptor Interacting Protein 1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(RIP1) Kinase Inhibitors; Analysis of Structure–
Kinetic Relationships
†
†
†
†
‡
Masato Yoshikawa, Morihisa Saitoh, Taisuke Katoh, Tomohiro Seki, Simone V. Bigi, Yuji
†
†
†
†
†
†
Shimizu, Tsuyoshi Ishii, Takuro Okai, Masako Kuno, Harumi Hattori, Etsuro Watanabe, Kumar
‡
‡
†
†
†
S. Saikatendu, Hua Zou, Masanori Nakakariya, Takayuki Tatamiya, Yoshihisa Nakada,
†
Takatoshi Yogo*
†Research, Takeda Pharmaceutical Company Limited, 26ꢀ1 MuraokaꢀHigashi 2ꢀchome, Fujisawa,
Kanagawa 251ꢀ8555, Japan
‡
Takeda Pharmaceuticals, 10410 Science Center Drive, San Diego, CA 92121, USA
KEYWORDS
Receptor interacting protein 1, multiple sclerosis, experimental autoimmune encephalomyelitis,
structure–kinetic relationship
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 110
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Abstract
We report the discovery of 7ꢀoxoꢀ2,4,5,7ꢀtetrahydroꢀ6Hꢀpyrazolo[3,4ꢀc]pyridine derivatives as a
novel class of receptor interacting protein 1 (RIP1) kinase inhibitors. Based on the overlay study
between HTS hit 10 and GSK2982772 (6) in RIP1 kinase, we designed and synthesized a novel class
of RIP1 kinase inhibitor 11, possessing moderate RIP1 kinase inhibitory activity and Pꢀgp mediated
efflux. The optimization of the core structure, the exploration of appropriate substituents utilizing
SBDD approach led to the discovery of 22, a highly potent, orally available, and brainꢀpenetrating
RIP1 kinase inhibitor with excellent PK profiles. Compound 22 significantly suppressed necroptotic
cell death both in mouse and human cells. Oral administration of 22 (10 mg/kg, bid) attenuated
disease progression in the mouse experimental autoimmune encephalomyelitis (EAE) model of
multiple sclerosis (MS). Moreover, analysis of structure–kinetic relationship (SKR) for our novel
chemical series was also discussed.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Page 3 of 110
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
INTRODUCTION
A process of programmed necrosis, termed necroptosis, has recently been recognized as an
1
important physiological process that involves cell death, innate immunity, and inflammation.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Receptor interacting protein 1 (RIP1) kinase is a crucial upstream regulator of necroptosis, and
association of RIP1 kinase with a variety of pathologies, such as ischemic injury, inflammatory
2
,3
diseases, and neurodegenerative diseases, has been investigated.
4
5
Neurodegenerative diseases, such as multiple sclerosis (MS), Parkinson's disease (PD), and
6, 7
amyotrophic lateral sclerosis (ALS),
have recently been intensively studied to understand their
relationship with the necroptosis pathway. Ofengeim et al. showed the involvement of RIP1 kinase
and necroptosis in mediating the deleterious process in MS by using both animal models and human
4
tissues from MS patients. RIP1ꢀmediated axonal degradation has also been demonstrated in ALS.
Thus, inhibitors of RIP1 kinase have the potential to provide new therapeutic opportunities to
neurodegenerative diseases with significant unmet medical needs.
8,9
The first published RIP1 kinase inhibitor, Necꢀ1 (1), was discovered by Degterev et al. in 2005,
10
with a subsequent disclosure of an analog, Necꢀ1s (2), with improved metabolic stability. Both
have been widely used to elucidate the biology of the necroptosis pathway. Over the recent years,
additional series of RIP1 kinase inhibitors have been identified as illustrated in Figure 1 such as a
11
12
13
typeꢀII RIP1 kinase inhibitor 3 and a hybridꢀtype RIP1 kinase inhibitor 4 which was designed
1
4
15
based on 2 and ponatinib. Among them, benzo[b][1,4]oxazepinꢀ4ꢀone analogs, represented by 5,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 110
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
6
comprise the most advanced series, with the optimized analog 6 currently being evaluated in
human clinical trials for inflammatory diseases such as ulcerative colitis, rheumatoid arthritis, and
psoriasis. However, despite the attractiveness of RIP1 kinase inhibitors as a therapeutic option for
neurodegenerative disorders, the chemical series reported so far are not suitable for clinical
development to treat CNS related disorders because of either low oral bioavailability and/or low
brain permeability. For example, the necrostatin series (2) has shown efficacy in animal models of
MS or ALS by oral administration, but its low metabolic stability has likely hindered its clinical
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
7
development. Compound 8 has been detected in mouse brains but only after intravenous injection.
18
The brain penetration or the in vivo efficacy of the other compounds (3ꢀ7) in neurodegenerative
diseases have not been reported to date. Thus, to develop RIP1 kinase inhibitors targeting CNS
diseases, we focused on the development of a novel chemical series of RIP1 kinase inhibitors with
good oral bioavailability and brainꢀpenetrating properties for treatment of neurodegenerative
diseases.
Herein,
we
report
the
discovery
of
(
3S)ꢀ3ꢀ(2ꢀbenzylꢀ3ꢀchloroꢀ7ꢀoxoꢀ2,4,5,7ꢀtetrahydroꢀ6Hꢀpyrazolo[3,4ꢀc]pyridinꢀ6ꢀyl)ꢀ5ꢀmethylꢀ4ꢀoxo
ꢀ2,3,4,5ꢀtetrahydroꢀ1,5ꢀbenzoxazepineꢀ8ꢀcarbonitrile (22) as a potent, orally bioavailable, and
brainꢀpenetrating RIP1 kinase inhibitor. The following strategies were used to identify 22: (i)
optimization of core structures for appropriate brain penetration in light of favorable
1
9
physicochemical properties such as topological polar surface area (TPSA) and the number of
ACS Paragon Plus Environment

Page 5 of 110
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
hydrogenꢀbonding donors (HBDs); (ii) exploration of potent substituents in back pockets by utilizing
structureꢀbased drug design (SBDD); and (iii) introduction of substituents toward the
solventꢀaccessible region to control the favorable ADMET profiles. Compound 22 exhibited potent
cellular activity both in humans and mice and an excellent pharmacokinetic (PK) profile with brain
exposure in mice after oral administration. Compound 22 was efficacious in experimental
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
20
autoimmune encephalomyelitis (EAE), which is a model of human inflammatory demyelinating
disease such as MS. Moreover, in the course of optimization, kinetic profiles of all synthesized
compounds were evaluated to analyze structure−kinetic relationships (SKRs). We also discuss the
SKR of our novel chemical series of RIP1 kinase inhibitors later in this paper.
Figure 1. Chemical structures of reported RIP1 kinase inhibitors.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 110
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
RESULTS AND DISCUSSION
Assay Development. The kinetics of the interaction between a ligand and its target has been
21,22
recognized as an important factor for lead generation/optimization
and long drugꢀtarget residence
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
3–25
time is often correlated with potent in vivo efficacy.
In the case of RIP1 kinase inhibitors, the
16
benzoxazepinoneꢀbased compounds have a long drugꢀtarget residence time (t1/2 = 112 min for 6),
which could translate into prolonged efficacy in cellular assays and potentially in vivo. These results
led us to analyze kinetic profiles including t1/2 in the course of the optimization to identity potent and
brainꢀpenetrating RIP1 kinase inhibitors. Measurement of the binding affinity for RIP1 kinase and
kinetic profiles was performed by using a timeꢀresolved fluorescence resonance energy transfer
(TRꢀFRET)
assay
and
a
novel
fluorescent
probe,
3ꢀ(3ꢀ((3ꢀ(4ꢀaminoꢀ5ꢀ(4ꢀ(3ꢀ(2ꢀfluoroꢀ5ꢀ(trifluoromethyl)phenyl)ureido)phenyl)ꢀ7Hꢀpyrrolo[2,3ꢀd]pyri
midinꢀ7ꢀyl)propyl)amino)ꢀ3ꢀoxopropyl)ꢀ5,5ꢀdifluoroꢀ7,9ꢀdimethylꢀ5Hꢀdipyrrolo[1,2ꢀc:2',1'ꢀf][1,3,2]d
iazaborininꢀ4ꢀiumꢀ5ꢀuide (9) described in Figure 2 (the synthesis of 9 is described in a
supplementary information). By using the TRꢀFRET assay, a highꢀthroughput assay format was
i
achieved sufficiently to evaluate all synthesized compounds. The binding affinity (pK ) was
calculated as the sum of logkon and pkoff. The drugꢀtarget residence time (t1/2) was defined as ln2/koff
.
Regarding compounds that quickly dissociate from the target (t1/2 < 5 min), the kinetic profiles could
not be precisely evaluated because of the detection limit of this TRꢀFRET assay.
ACS Paragon Plus Environment

Page 7 of 110
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2. Structure of fluorescent probe 9.
Lead Generation. To identify good starting points for lead optimization and subsequent
development, we performed highꢀthroughput screening (HTS) against RIP1 kinase by using our
i
compound library and identified a benzimidazole derivative 10 (pK = 7.20) (Figure 3A). Reported
RIP1 kinase inhibitors can be classified into three categories based on the binding modes; (i) typeꢀII
11
class targeting “DFGꢀout (or DLGꢀout in the case of RIP1 kinase)” conformation of kinase (e.g., 3),
11
(
ii) typeꢀIII (allosteric) class (e.g., 1, benzo[b][1,4]oxazepinꢀ4ꢀone series 5 and 6), or (iii) hybrid
13
type 4 (Figure 1). To clarify the binding mode of 10, an Xꢀray coꢀcrystal structure of 10 with RIP1
kinase was obtained (Figure 3B). The Xꢀray coꢀcrystal structure revealed that compound 10 had no
hingeꢀbinding interaction and occupied an allosteric lipophilic pocket at the back of the ATP binding
site, making a hydrogenꢀbonding interaction with the amide NH of Asp156 at a distance of 2.8 Å,
which indicated that the binding mode of compound 10 was typeꢀIII class similar to that of the Necꢀ1
and benzo[b][1,4]oxazepinꢀ4ꢀone series. The overlaid image of 10 and the
benzo[b][1,4]oxazepinꢀ4ꢀone derivative 6 in RIP1 kinase showed that the compounds occupied the
same space and that the benzyl group and the 5ꢀmembered heteroaromatic ring of compounds 6 and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 110
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
0 were wellꢀoverlapped (Figure 3C). This result led us to design and synthesize a novel class of
RIP1 kinase inhibitor 11, possessing a 4ꢀoxoꢀ6,7ꢀdihydroꢀ1Hꢀimidazo[4,5ꢀc]pyridine moiety as a
core structure (Figure 4). As a result, compound 11 showed moderate RIP1 kinase inhibitory activity
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(pK
i
= 6.93). Parent compound 6 had potent RIP1 kinase inhibitory activity with excellent in vitro
and in vivo profiles with highly favorable physicochemical and ADMET properties. However, 6 was
16
reported to have low brain penetration, presumably due to active extrusion of drug from the brain
via the efflux drug transporter Pꢀglycoprotein (Pꢀgp). On the other hand, compound 11 showed a
better Pꢀgp–mediated efflux (ER = 1.9) than that of 6 (ER = 3.1), likely reflecting the decreased
26,27
TPSA and HBD.
On the basis of this result, further 6,5ꢀbicyclic core structures were investigated
as replacements of the 4ꢀoxoꢀ6,7ꢀdihydroꢀ1Hꢀimidazo[4,5ꢀc]pyridine maintaining the favorable
2
28,29
physicochemical parameters (TPSA < 90 Å , HBD = 0) for brain exposure
and enhancing RIP1
kinase inhibitory activity.
ACS Paragon Plus Environment

Page 9 of 110
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(
A)
H
N
N
3
O N
2
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
0: pK = 7.20
i
(B)
(C)
Figure 3. HTS hit compound 10. (A) Structure of compound 10 and K
i
for human RIP1 kinase. (B)
Coꢀcrystal structure of 10 with hRIP1 kinase. (C) Overlay study between compounds 6 (yellow) and
1
0 (white) in hRIP1 kinase.
H
N
N
O N
2
H
N
O
O
0: pK = 7.20a
TPSAb = 72 Å , HBDb = 1
O
O
1
i
N
core
N
N
2
N
N
O
O
N
O
O
HN
H
a
1
1: pK = 6.93 , ER = 1.9
i
N
2
N
N
2
TPSA = <90 Å , HBD = 0
TPSA = 79 Å , HBD = 1
O
a
c
6
: pK = 9.01 , ER = 3.1
i
2
TPSA = 100 Å , HBD = 2
Figure 4. Design concept for a new class of brainꢀpenetrating RIP1 kinase inhibitors. (a) Binding
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 110
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
affinity for human RIP1 kinase. (b) Calculated by Daylight. (c) Efflux ratio in Pꢀgp–overexpressing
cells at 1 ꢁM.
The results of modification of the core structure are summarized in Table 1. Triazole derivative 12,
which was synthesized from 6 and removed HBD, showed dramatically decreased Pꢀgp–mediated
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
i
efflux and enhanced binding affinity (pK ) compared to those of 11. Although imidazole derivative
13, which has a nitrogen atom on a bridgehead, maintained moderate potency, 13 exhibited a higher
ER than that of 12 despite reducing TPSA, presumably due to the basicity of the imidazole moiety.
Pyrazole derivatives 14 and 15 also had a lower ER. These results in Table 1 indicated that our
design strategy of lowering TPSA and removing HBD was effective in reducing ER. Compound 14
showed the strongest binding affinity for RIP1 kinase and longest drugꢀtarget residence time.
Accordingly, the 7ꢀoxoꢀ2,4,5,7ꢀtetrahydroꢀ6
i
Hꢀpyrazolo[3,4ꢀc]pyridine moiety of 14 (pK = 8.41, t1/2
=
59 min, ER = 0.7) was identified as the most suitable core structure for further optimization.
ACS Paragon Plus Environment

Page 11 of 110
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 1. Modification of Core Structure
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
c,d
TPSA
t1/2
b
c
c
Compd
core
HBD ER
pK
i
logkon
pkoff
2
(Å )
(min)
6.27
0.984
6.7
1
1
1
2
3
4
81
68
68
68
0
0
0
0
0.8 7.26
1.2 6.97
0.7 8.41
0.49 6.60
(6.04–6.43) (0.803–1.30) (4.4–14)
5.85
1.13
9.3
(
(
5.55–6.02) (0.909–1.58) (5.6–26)
6.48
1.93
59
6.39–6.55) (1.83–2.06)
(46–80)
e
e
e
e
15
11
n.d.
n.d.
n.d.
n.d.
<5
<5
79
1
1.9 6.93
a
b
c
Calculated by Daylight. Efflux ratios in Pꢀgp–overexpressing cells at 1 ꢁM. 95% confidence
d
e
intervals (given in parentheses) determined in duplicate (n = 2). t1/2 = ln2/koff. not determined.
Lead Optimization. A docking model of 14 with RIP1 kinase was developed to utilize SBDD for
further optimization (Figure 5A). This model showed that the benzyl group deeply occupied the
allosteric hydrophobic pocket (we refer to this pocket as deep pocket 1: DP1) at the back of the ATP
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 110
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
binding site. The DP1 was mainly constructed by the side chains of lipophilic amino acid residues
such as Leu129 and Val134. According to the Xꢀray coꢀcrystal structure of 2 with RIP1 kinase, a
hydrogenꢀbonding interaction was observed between the OH of Ser161 and NH of the indole in 2 in
DP1. As shown in Figure 5B, a small unoccupied pocket, which is defined by Met67 and Leu70, was
found around the 3ꢀposition of the core structure (referred to as deep pocket 2: DP2). The docking
model also suggested that the 7ꢀ and 8ꢀpositions of the benzoxazepinone ring were oriented toward
the solventꢀaccessible region. On the basis of these analyses, optimization of 14 was continued by (i)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
investigation of favorable substituents in DP1 (R ) in view of the lipophilicity, size, or hydrogen
2
bonding with OH of Ser161, (ii) exploration of substituents in DP2 (R ) that can efficiently occupy
3
the small lipophilic pocket, and (iii) introduction of an appropriate substituent (R ) into the 7ꢀ or
8ꢀpositions of the benzoxazepinone ring to control the favorable ADMET profiles (Figure 5C).
ACS Paragon Plus Environment

Page 13 of 110
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(
A)
(B)
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(C)
Met67
DP2
Leu78
(
ii)
Leu90
R2
3
O
N
Ser161
Val134
N
N
R1
N
Lys45
(i)
O
7
O
DP1
iii) R3
8
Leu129
(
His136
solvent-accessible region
Figure 5. Synthetic strategy based on the docking model. (A) The docking study between 14 and
RIP1 kinase. The DLG motif including Asp156 in the activation loop was omitted for simplicity. (B)
Image of the protein surface around DP2. (C) Schematic illustration of 14 in RIP1 kinase.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 110
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
First, modification of the substituent R was conducted. The results are summarized in Table 2 (The
detailed SKR analysis for all synthesized compounds will be discussed (vide infra)). Introduction of
a fluoro group into the
introduction into the ꢀ and
of the nonꢀsubstituted analog 14. An additional fluoro group was introduced into the second
ꢀposition of 16a, yielding compound 16d, increasing the binding affinity. Replacement of the
o
ꢀposition resulted in enhanced binding affinity (compound 16a), whereas
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
m
p
ꢀpositions (compounds 16b,c) decreased the affinity compared to that
o
phenyl group with a cyclobutyl group resulted in reduced binding affinity. To direct a
hydrogenꢀbonding interaction toward the OH of Ser161, substituents possessing a hydrogenꢀbonding
acceptor (HBA), such as heteroaromatic rings and sulfone groups, were introduced. Pyridine analogs
1
i
6f and 16g exhibited 10ꢀfold lower K values than that of 14, whereas introduction of sulfone
groups (16h and 16i) dramatically decreased the binding affinity. Among all compounds with
1
modification of substituents for R , the 2ꢀfluorophenyl analog 16a and 2,6ꢀdifluorophenyl analog
1
6d showed more potent binding affinities and longer drugꢀtarget residence times. Therefore,
1
2
ꢀfluorophenyl and 2,6ꢀdifluorophenyl groups were selected as favorable substituents for R for
further optimization.
ACS Paragon Plus Environment

Page 15 of 110
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
Table 2. Modification of Substituent R
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a,b
1/2
t
1
a
a
Compd
R
pK
i
logkon
pkoff
(
min)
110
6.49–6.59) (2.14–2.30) (96–140)
6.54
2.21
16a
8.76
8.29
8.10
8.94
7.86
7.47
7.42
5.16
5.37
8.41
(
(
(
(
6.48
6.40–6.54) (1.73–1.94)
6.65 1.45
6.52–6.75) (1.32–1.65)
.58 2.36
1.82
46
(37–60)
20
1
6b
6c
6d
6e
1
(14–31)
160
6
1
6.54–6.61) (2.31–2.43) (140–180)
6
.49
(6.33–6.60) (1.22–1.61)
.90 1.58
5.77–6.00) (1.45–1.76)
.98 1.44
5.81–6.10) (1.28–1.69)
1.37
16
(11–29)
26
1
5
16f
(
(
(19–40)
19
5
16g
(13–34)
c
c
16h
n.d.
n.d.
n.d.
n.d.
<5
<5
c
c
16i
6.48
1.93
59
14
(6.39–6.55) (1.83–2.06)
(46–80)
a
b
c
9
5% confidence intervals (given in parentheses) determined in duplicate (n = 2). t1 = ln2/k . not
/2 off
determined.
2
Second, the results of introduction of the substituent R are summarized in Table 3. The methyl
i
analog 17 showed slightly enhanced pK compared to that of the nonꢀsubstituted derivative 14.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 110
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Introduction of a chloro group (18) significantly boosted the binding affinity and resulted in marked
prolongation of t1/2 (t1/2 = 210 min). Compounds 19 and 20, which had polar substituents such as
cyano and carboxamide groups, also enhanced their binding affinity. These results demonstrated that
filling the DP2 with not only lipophilic groups but also polar groups, was remarkably effective to
enhance the binding affinity. Accordingly, a chloro group was identified as the most potent and
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
i
slowꢀdissociating substituent (pK = 9.39, t1/2 = 210 min).
Table 3. Effect of Introducing a Substituent into the 3ꢀPosition of the Core Structure
a,b
1/2
t
2
a
a
Compd
R
pK
i
logkon
pkoff
(
min)
130
(6.25–6.35) (2.21–2.39) (110–170)
.91 2.48 210
6.87–6.95) (2.41–2.57) (180–260)
.52 2.42 180
6.46–6.57) (2.33–2.55) (150–240)
6
.30
2.29
1
1
7
8
Me
Cl
8.59
9.39
8.94
6
(
(
6
19
CN
6
.27
2.33
150
2
0
8.60
8.41
(
(
6.21–6.32) (2.24–2.44) (120–190)
6.48
1.93
59
14
H
6.39–6.55) (1.83–2.06)
(46–80)
a
b
95% confidence intervals (given in parentheses) determined in duplicate (n = 2). t1/2 = ln2/koff
.
ACS Paragon Plus Environment

Page 17 of 110
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
2
We identified the favorable substituents for R and R to show potent binding affinity and slow
1
2
dissociation (i.e., 2,6ꢀdifluorophenyl and 2ꢀfluorophenyl groups for R and chloro group for R ), so
further optimization of the scaffold was continued to control the favorable ADMET profiles by using
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
substituent R toward the solventꢀaccessible region. Compounds 16a, 16d, and 18 bearing these
i
optimal substituents showed potent pK and long t1/2 value, but they had an issue of poor metabolic
stabilities (data for 18 in Table 4, and data for 16a and 16d in a supplementary information). We
assumed the benzene ring of the benzoxazepinone would be metabolically labile due to the high
electron density. Therefore, electronꢀwithdrawing and polar substituents such as cyano and amide
3
groups were introduced as R into the 7ꢀ and 8ꢀpositions of the benzoxazepinone ring to block the
assumed metabolic site, reduce the electron density of the benzoxazepinone, and decrease the total
3
lipophilicity of the molecule. The results of introduction of the substituent R and the combination
1
2
between R and R are summarized in Table 4. Introduction of cyano groups into the 7ꢀ and
8
ꢀpositions improved the metabolic stabilities while maintaining potent binding affinity. In particular,
the 8ꢀcyano analog 22 showed better metabolic stability than that of the 7ꢀcyano analog 21. This
result suggests that a metabolic soft spot could be the 8ꢀposition of the benzoxazepinone. Amide
groups were introduced into the 8ꢀposition instead of the cyano group of 22. The amide analogs 23
and 24 showed high Pꢀgp–mediated efflux (ER = 17 and 15, respectively) despite maintaining
moderate to potent binding affinity and satisfactory metabolic stability. Compounds 25 and 26
possessing 2ꢀfluorophenyl or 2,6ꢀdifluorophenyl groups were synthesized. Compounds 25 and 26
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 110
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
had moderate potency and good metabolic stability in addition to improved Pꢀgp–mediated efflux
(ER = 1.1 and 1.3, respectively). On the basis of these results, we selected compound 22 for further
evaluation considering the balance between potency, metabolic stability, and Pꢀgp ER. We also
evaluated the selectivity of compound 22 for RIP1 kinase by using Reaction Biology Corp kinase
panel and Eurofins Panlabs panel. As a result, compound 22 represented excellent specificity for
RIP1 kinase over 406 kinases including RIP3 kinase (%inhibition: <50% at 10 ꢁM) other than
LIMK2 (>50% at 10 ꢁM) in Reaction Biology Corp kinase panel. Compound 22 also showed
excellent selectivity against 106 targets in Eurofins Panlabs panel. The results of Reaction Biology
Corp kinase panel and Eurofins Panlabs panel are listed in supporting information (Supplemental
Tables 4 and 5).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Page 19 of 110
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
2
3
Table 4. Optimization of Substituents R , R , and R
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a,b
1/2
c
t
HLM/MLM
1
2
3
a
a
d
Compd
R
R
R
pK
i
logkon
pkoff
ER
(
min) (ꢁL/min/mg)
6
.58
2.49
(2.43–
2.55)
2.49
210
(190–
250)
210
2
1
Cl
Cl
7ꢀCN
8ꢀCN
9.07
9.04
(6.55–
6.61)
27/38
<1/<1
13/<1
19/7.0
<1/<1
<1/11
0.7
6
.55
(6.51–
.59)
.52
(6.49–
.54)
22
(2.41–
2.58)
2.52
(180–
260)
230
0.6
17
6
6
23
24
25
26
18
Cl 8ꢀCONHMe 9.04
(2.47–
2.57)
2.32
(200–
260)
150
6
6.13
Cl 8ꢀCONMe
2
8.46
8.53
8.75
9.39
(6.09–
(2.27–
2.39)
2.33
(130–
170)
150
15
6
.17)
6
.19
H
H
8ꢀCN
8ꢀCN
H
(6.15–
6.24)
(2.26–
2.42)
2.47
(130–
180)
200
1.1
6
.29
(6.23–
.33)
.91
(6.87–
.95)
(2.38–
2.58)
2.48
(170–
260)
210
1.3
6
6
e
Cl
(2.41–
2.57)
(180–
260)
101/100
NT
6
a
c
b
95% confidence intervals (given in parentheses) determined in duplicate (n = 2). t1 = ln2/koff.
/2
d
Metabolic stability in human and mouse liver microsomes. Efflux ratios in Pꢀgp–overexpressing
e
cells at 1 ꢁM. Not tested.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 110
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Crystal Structure of Compound 22. An Xꢀray coꢀcrystal structure of 22 bound to RIP1 kinase is
shown in Figure 6. As we expected, the type III binding mode was observed. The core structure
formed a hydrogenꢀbonding network with Asp156 among the nitrogen atom at the 1ꢀposition and the
carbonyl group at the 7ꢀposition. The benzyl group fit into the allosteric hydrophobic DP1 at the
back of the ATP binding site. The chloro group effectively occupied the small DP2, presumably
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
0,31
interacting with the carbonyl group of Met67 at a distance of 3.5 Å.
This chloroꢀoxygen
interaction might result in extremely high affinity of 18 compared to that of the methyl analog 17.
The cyano group on the benzoxazepinone ring was exposed toward the solventꢀaccessible region. No
interaction was observed between 22 and Ser161 which is an autophosphorylation site. Reflecting
this type III binding mode, compound 22 represented excellent selectivity against 406 kinases
including RIP3 kinase (%inhibition: <50% at 10 ꢁM) other than LIMK2 (>50% at 10 ꢁM). These
results demonstrated that compound 22 would be an excellent tool compound for in vitro and in vivo
experiments.
ACS Paragon Plus Environment

Page 21 of 110
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 6. Xꢀray crystal structure of compound 22 (white).
In Vitro and In Vivo Profiling of Selected Candidate. Compound 22 was tested further in an in
vitro binding assay for human and mouse RIP1 kinase proteins (Table 5). Compound 22 had
approximately 90 times more potent affinity for human RIP1 kinase than that for mouse RIP1 kinase
in TRꢀFRET assay. Cellular assays were also performed by using human colorectal adenocarcinoma
HTꢀ29 cells and mouse Lꢀcells NCTC 929 (L929). L929 necroptosis assay has been generally used to
32
examine the necroptosis pathways in mouse cells and regarded as wellꢀestablished cellular model.
Compound 22 strongly suppressed necroptotic cell death and phosphorylation of MLKL (pMLKL),
which was one of the hallmark mediators of necroptosis, in HTꢀ29 cells (Table 5) (nectoptosis; IC =
5
0
2
.0 nM, pMLKL; IC = 1.3 nM) as well as L929 cells (nectoptosis; IC = 15 nM, pMLKL; IC =
50 50 50
2.7 nM). The high correlation was observed between TRꢀFRET and cellular assays in human,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 110
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
whereas there was a discrepancy between TRꢀFRET and cellular assays in mouse compared to
human. It was not clear why there was a bigger discrepancy between TRꢀFRET and cellular assay of
compound 22 in mouse than that in human, however the similar trend was observed between human
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
33
and mouse in the other 7ꢀoxoꢀ2,4,5,7ꢀtetrahydroꢀ6Hꢀpyrazolo[3,4ꢀc]pyridine derivatives and
1
6
compound 6. In the cellular conditions, endogenous fullꢀlength enzyme of RIP1 kinase is expressed
34–36
and forms a complex with a variety of proteins.
On the other hand, RIP1 kinase in the TRꢀFRET
assay was composed of the catalytic domain of RIP1 kinase without any associated proteins.
Although there is no information about the structural difference between the truncated catalytic
domain of RIP1 kinase and the intracellular intact RIP1 kinase, the structural difference between
truncated enzymes of human and mouse RIP1 kinase might explain the discrepancy between
TRꢀFRET and cellular assays in mouse compared to human. Furthermore, in the case of L929 cells,
TNFꢀα prominently causes necrosis instead of apoptosis, and this response is different from the other
37
cell lines. In fact, HTꢀ29 cells require TNFꢀα, cIAP inhibitor, and a panꢀcaspase inhibitor to induce
32
necroptosis. Therefore, the necroptosisꢀinducible conditions are highly dependent on the kind of
cell line. The L929 cells might be sensitive to necroptosis inducer and not require full target
occupation of endogenous RIP1 kinase. Considering the excellent specificity for RIP1 kinase of 22
3
3
and the result of in vivo target engagement in mouse using cellular thermal shift assay (CETSA) , it
is indicated that cellular activities of 22 should be caused by the inhibition of RIP1 kinase.
Compound 22 was progressed to an in vivo study to evaluate its efficacy through inhibition of
ACS Paragon Plus Environment

Page 23 of 110
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
nonꢀprimate RIP1 kinase.
Table 5. Cellular activities in HTꢀ29 and L929 cells
Necroptosis
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
pMLKL
IC (nM)
pK
i
IC (nM)
Compd
50
50
a,b
a,c
a
a,b
a,c
human
mouse
human
mouse
human
mouse
2
2
9.04 7.09 (6.96–7.23) 2.0 (1.5–2.7) 15 (14–17) 1.3 (0.3–5.5) 2.7 (1.9–3.7)
a
c
b
95% confidence intervals (given in parentheses) determined in duplicate (n = 2). HTꢀ29 cells.
L929 cells.
PK study of compound 22 was conducted and the results are shown in Table 6. Compound 22 was
administered to mice, and the plasma concentration was determined after oral (po, 1 mg/kg) and
intravenous (iv, 0.1 mg/kg) administration. Compound 22 exhibited a desirable PK profile including
high plasma exposure (AUC = 658 ng·h/mL) and a moderate plasma duration (MRT = 3.1 h).
a
Table 6. PK Parameters of Compound 22 in Mice
PK parameters in mouse
Compd
CL*
mL/h/kg) (ng/mL
794 140.1
C
max** MRT** AUC0ꢀ8h**
F
(
)
(h)
3.1
(ng·h/mL) (%)
658 52.2
22
a
The values are shown as means of three determinations (0.1 mg/kg, iv*, and 1.0 mg/kg, po**).
Moreover, compound 22 showed satisfactory brain exposure (Cb,u = 0.0230 ꢁg/g), which was
equivalent to 50 nM, and a brainꢀtoꢀplasma ratio (Kp,uu = 0.3) 1 hour after oral administration (po,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 110
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
0 mg/kg) in mice (Table 7). We have also confirmed the target association of 22 in the mouse brain
3
3
tissues by using CETSA at 1 hour after oral administration of 22. On the basis of these results,
compound 22 was selected for further evaluation in the EAE models of MS.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
Table 7. Concentrations in Plasma and Brain of Compound 22 in Mice
time
Conc. (ꢁg/mL or ꢁg/g)
f
g
Compd
Kp
0.7
Kp,uu
0.3
b
c
d
e
(h)
Cp
Cb
Cp,u
Cb,u
22
1
1.479
1.038
0.089
0.0230
a
d
b
c
Administered orally at 10 mg/kg (n = 3). Total plasma concentration. Total brain concentration.
e
f
g
Unbound plasma concentration. Unbound brain concentration. Kp = Cb/Cp. Kp,uu = Cb,u/Cp,u.
Compound 22 was orally administered from day 0 to day 26 and progression of clinical signs of
EAE was daily monitored. Clinical signs of EAE in vehicleꢀtreated mice were observed from day 12,
and peaked at around day 21. On the other hand, the orally administered RIP1 kinase inhibitor 22 at
doses of 10 mg/kg twice a day (20 mg/kg/day) attenuated the development of clinical symptoms of
EAE (Figure 7A). Cumulative score of EAE was significantly lower than that of the vehicleꢀtreated
group (Figure 7B). These results indicated that compound 22 protected against disease progression in
the EAE model. Ofengeim et al. also reported that Necꢀ1s (7Nꢀ1, compound 2 in this paper) blocks
4
the disease progression in the EAE model. Although it is difficult to compare the efficacy between
22 and 2 without headꢀtoꢀhead comparison in the same condition, compound 22 would be a novel
class of an orally available and brainꢀpenetrating in vivo tool for RIP1 kinase.
ACS Paragon Plus Environment

Page 25 of 110
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 7. Assessment of compound 22 in the EAE model.
RIP1 kinase inhibitor 22 (10 mg/kg) or vehicle (0.5% methylcellulose) was orally administered to
mice twice a day for 26 days. (A–B) The EAE clinical score of (A) the time course and (B) the
cumulative clinical score from day 0 to day 26 are shown. Data are presented as the mean ± S.E. (n =
12). **p < 0.05 compared to the naïve group (Dunn's test).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 110
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Structure−kinetic relationship.
Several properties of ligands, including molecular weight, lipophilicity, and rotatable bond count,
38
have been reported to affect kinetic parameters.
During the optimization of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7ꢀoxoꢀ2,4,5,7ꢀtetrahydroꢀ6Hꢀpyrazolo[3,4ꢀc]pyridine derivatives, molecular weight and rotatable
bond count were not changed greatly and lipophilicity was the main parameter to optimize profiles
1
3
such as binding activity and metabolic stability. Therefore, the SKR at R –R substituents were
analyzed in relation with lipophilicity. We also investigated the relationship between kon and koff to
know the value of SKR analysis compared to that of conventional SAR.
1
In the modification of the substituent for R , relationships between kinetic profiles (logkon: blue
dots, pkoff: red dots) and lipophilicity are summarized in Figure 8 (the detailed kinetic parameters are
listed in a supplementary information). Introduction of a fluoro group into the benzene ring in DP1
affected the pkoff more than the logkon (14 and 16a–d). Replacement of phenyl group with cyclobutyl
group maintained the logkon value and resulted in a 4ꢀfold decrease in the koff value (14 and 16e).
This finding suggested that the decreased koff value of 16e caused the decreased binding affinity. The
pyridine analogs 16f and 16g exhibited slightly lower logkon and pkoff values than that of 14, whereas
introduction of sulfone groups (16h and 16i) dramatically decreased both logkon and pkoff values.
Overall, the logkon and pkoff values were mostly raised depending on the degree of increase in the
clog P, and a particularly good correlation was observed between the logkon and clog P.
ACS Paragon Plus Environment

Page 27 of 110
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
Figure 8. SKR and relationships between kinetic profiles and lipophilicity in the modification of R .
1
Relationship between the logk and pk in modification of the substituent at R is displayed in
on
off
Figure 9. No apparent correlation was observed between the logk and pk . Interestingly, it was
on
off
revealed that the k and k changed independently although both the logk and pk values roughly
on
off
on
off
correlated with clog P. These results suggested that k and k can be optimized independently by
on
off
measuring SKR, which would provide another medicinal chemistry strategy to obtain more potent
compounds by combining the best k moiety with the best k moiety.
on
off
2
3
2
Regarding the substituents at R and R , introduction of the appropriate R substituent resulted in a
longer drugꢀtarget residence time, but no notable effect on kinetic profiles was observed for
3
substituent R (Tables 3 and 4).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 110
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
Figure 9. Relationships between logkon and pkoff in modification of R .
The cellular activity and binding parameters are summarized in Table 7. Based on the cocrystal
structure of 22, compounds 14 and 22 should be type III inhibitors and would also be ATP
16
competitive inhibitors as 3 and 6. In comparison between four ATP competitive inhibitors in Table
7
, slower dissociating compounds exhibited more potent cellular activity. The residence time (t )
1/2
was more correlated with the cellular activity than the affinity (pK ). It suggested that in the case of
i
RIP1 kinase inhibitors, the residence time would be more important for achieving optimal cellular
activity than affinity. However, further study is needed to conclude which is more critical for the
cellular activity, the affinity or the residence time, because the sample size is small in this study. By
ACS Paragon Plus Environment

Page 29 of 110
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
analyzing kinetic profiles as well as thermodynamic profiles, we succeeded to identify a favorable
RIP1 kinase inhibitor that showed the potent cellular activity.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 7. Impact of t1/2 on cellular activity
b,c
HT29 Necroptosis
t
1/2
b
b
Compd pK
i
logkon
7.63
pkoff
a,b
IC50 (nM)
(min)
8.0
300
1.06
3
8.69
(
160–580)
18
(7.56–7.69) (0.99–1.15) (6.8–9.9)
6.48
(6.39–6.55) (1.83–2.06)
6.79 2.22
1.93
59
(46–80)
115
14
8.41
9.01
9.04
(14–23)
4.1
6
(
2.7–6.1)
.0
1.5–2.7)
(6.74–6.84) (2.14–2.31) (97–143)
6.55 2.49 210
(6.51–6.59) (2.41–2.58) (180–260)
2
22
(
a
b
c
HTꢀ29 cells. 95% confidence intervals (given in parentheses) determined in duplicate (n = 2). t1/2
=
ln2/koff
.
CONCLUSION
ꢀOxoꢀ2,4,5,7ꢀtetrahydroꢀ6Hꢀpyrazolo[3,4ꢀc]pyridine derivatives were discovered as potent and
7
brainꢀpenetrating RIP1 kinase inhibitors. Optimization of the core structure and exploration of
appropriate substituents for DP1, DP2, and solventꢀaccessible region by utilizing the SBDD
approach
led
to
the
discovery
of
(3S)ꢀ3ꢀ(2ꢀbenzylꢀ3ꢀchloroꢀ7ꢀoxoꢀ2,4,5,7ꢀtetrahydroꢀ6Hꢀpyrazolo[3,4ꢀc]pyridinꢀ6ꢀyl)ꢀ5ꢀmethylꢀ4ꢀoxo
ꢀ2,3,4,5ꢀtetrahydroꢀ1,5ꢀbenzoxazepineꢀ8ꢀcarbonitrile (22), a highly potent, orally active, and
brainꢀpenetrating RIP1 kinase inhibitor with an excellent PK profile. Compound 22 significantly
ACS Paragon Plus Environment